Page 39 - 32.3

Meanwhile, we also examine the information opacity of accounting information.
NTU Management Review Vol. 32 No. 3 Dec. 2022
Although we propose that the volume of accounting information should be similar across

firms, the dissimilarity of information quality may lead to variations in information
firms, the dissimilarity of information quality may lead to variations in information
opacity across firms. To capture the information quality of accounting information, we
opacity across firms. To capture the information quality of accounting information, we
use discretionary accruals (DACC) as our proxy. Recognizing that there are two common
use discretionary accruals (DACC) as our proxy. Recognizing that there are two common
methods for calculating discretionary accruals (Dechow, Sloan, and Sweeney, 1995;
methods for calculating discretionary accruals (Dechow, Sloan, and Sweeney, 1995;
Kothari, Leone, and Wasley, 2005), we establish our regression models by using both.
Kothari, Leone, and Wasley, 2005), we establish our regression models by using both. The
The two regression models are specified as follows:
two regression models are specified as follows:

ln(TP i ,t ) = γ + γ 1 DACCa i ,t + γ 2 ROA + γ 3 RET + γ 4 DACCa i ,t × ROA i ,t
i
,t
i
0
,t
,t 
+ γ 5 DACCa i ,t × RET + γ k Control V ariables i ,t + ε i ,t , (3)
i
ln(TP i ,t ) = δ + δ 1 DACCb i ,t + δ 2 ROA + δ 3 RET + δ 4 DACCb i ,t × ROA i ,t (4)
i
,t
0
i
,t
,t 
+ δ 5 DACCb i ,t × RET + δ k Control V ariables i ,t + ε i ,t ,
i
DACCa as specified in equation (3) is calculated using the formula provided by
DACCa as specified in equation (3) is calculated using the formula provided by
Dechow et al. (1995), and DACCb as specified in equation (4) is calculated using the
Dechow et al. (1995), and DACCb as specified in equation (4) is calculated using the
formula provided by Kothari et al. (2005).
formula provided by Kothari et al. (2005).
We measure the information quality of accounting information using the absolute val-
We measure the information quality of accounting information using the absolute
ue of discretionary accruals. If the information quality of accounting information is low,
value of discretionary accruals. If the information quality of accounting information is
suggesting that the information opacity of accounting information is high, we expect the
low, suggesting that the information opacity of accounting information is high, we expect
weighting of the accounting-based performance measure to decrease and the weighting of
the weighting of the accounting-based performance measure to decrease and the
the stock-based performance measure to increase at the same time. Therefore, we predict
weighting of the stock-based performance measure to increase at the same time. Therefore,
the sign of γ4 and δ4 in equations (3) and (4) to be negative and the sign of γ5 and δ5 in equa-
tions (3) and (4) to be positive.
we predict the sign of γ4 and δ4 in equations (3) and (4) to be negative and the sign of γ5
and δ5 in equations (3) and (4) to be positive.
3. Findings

3. Findings
First, we test equations (1) and (2). The estimated coefficients of ROA × SPREAD and
RET × SPREAD are 1.858 (p-value = 0.002) and -0.105 (p-value = 0.064), respectively;
First, we test equations (1) and (2). The estimated coefficients of ROA × SPREAD
the estimated coefficients of ROA × ANALYSTS and RET × ANALYSTS are -0.033 (p-value
and RET × SPREAD are 1.858 (p-value = 0.002) and -0.105 (p-value = 0.064),
= 0.002) and 0.004 (p-value = 0.003), respectively. These results are consistent with our
respectively; the estimated coefficients of ROA × ANALYSTS and RET × ANALYSTS are
hypotheses, indicating that the weighting of the stock-based performance measure decreas-
-0.033 (p-value = 0.002) and 0.004 (p-value = 0.003), respectively. These results are
es significantly if the information opacity of the stock-related information is high, and vice
consistent with our hypotheses, indicating that the weighting of the stock-based
versa.

43 29